mailing-list for TeXmacs Users

[Lionel Elie Mamane <address@hidden>]

OK, let's see if I understand...

On Tue, Nov 30, 2004 at 11:47:34AM +0100, Joris van der Hoeven wrote:

> On Mon, 29 Nov 2004, Lionel Elie Mamane wrote:

>> The way X does modifiers has changed. This broke at least Emacs,
>> TeXmacs and sawfish (I think also the KDE wm); the X guys say that
>> these programs are broken. A "fix" for Emacs is now in the Debian
>> package.

> What seems to happen is that the windows key is interpreted as H-M
> (hyper-meta) due to a newly introduced bug in X

When I press "Flag-d" in xev, the state is "0x40 which means
(Appendix B of "X Window System Protocol"):

 - Mod4 on

But then mod4 is generated by the keysyms "Meta_L", Super_L, Hyper_L
and Meta_L (according to xmodmap), which means (section 6.6 of ICCCM)
that mod4 is Meta AND is Super AND is Hyper.

Right?

OK, so I do:

xmodmap -e 'remove mod4 = Hyper_L'
xmodmap -e 'remove mod4 = Super_L'

Now the flag key is recognised as M, but also as A... At least, M-'
accents the next character. Or maybe is M-' an alias for A-' in
TeXmacs?

> In any case, it think that this is a bug in X, since it seems
> unreasonable in the first place that one key corresponds to several
> modifiers.

If I understand it right, it is the (numbered: mod4) modifier that is
mapped to several (non-numbered: Meta, Super, Hyper) modifiers.

> That said, we might want to invent a fix for TeXmacs, since the bug
> seems to annoy several persons. However, this does require some
> work, because TeXmacs currently does not read the xmodmap at
> start-up,

How does it determine which numbered modifier is which named modifier,
then? Can you please give me a headstart and point me to the right
file? Thanks.

For what is worth, the ICCCM says one should... mod1 is Alt if and
only if Alt_L or Alt_R maps to it, is Meta if and only if Meta_L or
Meta_R maps to it, etc.

> If someone wants to hack the sources and adapt the Emacs patch, then
> I will apply the fix. However, I don't have much time right now and
> I cannot test any fixes myself, since I don't have this problem.

I'll gladly test any fix anybody makes.

-- 
Lionel